TikTok ads: attention is still cheap here, and it will not last

This platform works to a rule the other four do not have, and everything else follows from it: your followers count for almost nothing in distribution. Every video is shown to strangers, judged on its own results, and distributed or dropped within hours.

That inverts the advice given everywhere else in this series. There is no audience to assemble before starting: an account opened this morning can beat a three-year-old one on its first video, and it has happened to businesses we know.

It also inverts the economics of production. A video that looks like an advertisement is dismissed in a second; a person talking to their phone in their shop is not. Production value is a negative signal here, which is unpleasant for an agency to write.

Finally, there is a precise reason to be here now rather than in two years, and we write it with its expiry attached: attention is cheap because the audience has grown faster than the number of advertisers. That gap is what you are buying, and it is closing.

This is the most expensive sentence in this article for us, and it can be verified in an evening: videos that look like advertisements get dismissed, and videos that look like what people already watch do not.

Stock music, a drone shot, an animated logo, centred text in a smooth typeface: each of those tells the viewer "this is a paid message", and the gesture that follows is immediate and automatic.

What works instead fits inside a hand-held phone: a person speaking into the lens, in their workshop, with the noise of the place, saying something useful in the first few words. That is not carelessness, it is a register.

There is one exception worth knowing so you do not overcorrect: the picture has to be sharp and the sound audible. The register is amateur, legibility is not, and a video nobody can hear is dismissed as fast as an advertisement.

The budget consequence is the one no quotation likes to carry: on this platform an expensive shoot often produces worse results than an employee filming three things a week. We prefer to write that down than to discover it with you after a quarter.

On the other platforms, answering comments is a matter of reputation and conversion. Here a further function is added: it is a means of distribution.

A question asked in a comment can become the next video, and that format — a filmed answer to a real question — is among the most watched on the platform. You get a subject, a hook and a reason to exist, free, written by somebody else.

Comments also feed distribution directly: a video with comments is shown to more people. A question asked sincerely in the caption — "which one do you prefer?", "how many should I keep?" — raises that signal at no cost.

Moderation keeps the same rules as in the Facebook article: hide insults and numbers left by competitors, never a genuine question. Here a mechanical reason not to hide is added: every deleted comment is one distribution signal fewer.

And one habit that pays for itself: keep a list of the questions that have come back more than three times. It is simultaneously your publishing calendar, your list of common questions, and proof that your profile is not saying something it should.

Two platform figures are worth reading, and they are free. The full-watch rate, which says whether your video holds to the end — it is the first signal distribution looks at.

Saves and shares, as on Instagram: they describe an intention rather than an approval. A saved video is one somebody means to rewatch before buying, or to show to whoever pays.

Two figures do not deserve your attention: the view count, which depends mostly on the national distribution described above, and the follower count, which distributes nothing here and therefore measures nothing.

Your own figures remain those of this whole pillar: messages received, share handled within the hour, orders accepted, origin asked of the customer. The Facebook article sets them out and they apply unchanged.

And we will quote no cost per view or per message for Algeria: we know of no dated, checkable series. What we can say is that the gap described in the first section makes those costs lower here than elsewhere today, and that you should measure yours before they rise.

What we will refuse: producing videos for this platform that look like advertisements. It is what sells best and fails fastest, and saying so before the quotation regularly costs us the project.

We will refuse to fund a video that has not first been put into free distribution, and to republish here a file carrying another platform’s watermark. The first buys an audience with no evidence, the second pays to be seen less.

We will also refuse to propose this platform to a business whose trade cannot be shown — an administrative service, a trading operation with no workshop, an activity with nothing to film. No window of cheap attention makes up for having no subject.

What we do: a frequency rather than a campaign; the filming handed to somebody inside your business rather than a crew; the original files staying with you; paid distribution restricted to your area and reserved for videos already validated for free; and the number of messages from your own town as the judge.

And what you should do without us this week: film three twenty-second videos — a before and after, the unpacking of an order, and you saying a price out loud. Publish them without funding them, three days apart, and see which one holds. It costs nothing and you will know whether there is anything here for you.